7 
remarkable difference between the flowers of Collinsia and Schl- 
zanthus, which had at first escaped my attention. In both the 
stamina are apparently decimate, enclosed in the carinate middle 
division of the lower lip of the corolla, whilst the lateral lobes are 
expanded at right angles. In Collinsia this is really the case, 
but in Schizanthus the pedicel is twisted, the flower resupinate, 
and in relation to the parts of the flower the stamina are in fact 
ascending and enclosed in the carinate upper lip. The real place 
of these two genera still remains therefore to be decided on. 
The Digitaleae, consisting of five well-known genera, are chiefly 
distinguished from Gratiolese by the filaments of the stamina, which 
immediately above their insertion are bent downwards, although 
they usually are afterwards curved up, so as that the anthers are 
collected under the upper lip of the corolla. The capsule is also 
usually harder, and always deeply sulcate, but their real affinity to 
Gratiolese is very close. The genera Digitalis and Isoplexis have 
been described in detail in Lindley’s Monograph, illustrated by Bauer’s 
splendid drawings. Russelia is at once separated from Pentstemon 
by the absence of the fifth sterile stamen. Between Chelone and 
Pentstemon there has been much uncertainty, some authors consider- 
ing as Chelone all the species with hirsute anthers, and as Pentste- 
mon those in which they are glabrous ; others depending on the 
form of the corolla only ; others again combining that consideration 
with the presence or absence of the membranous wing round the 
seed. I have followed the latter opinion, referring to Chelone the 
old C. glabra, obliqua and Lyoni, and, as a somewhat anomalous 
species, the C. nemorosa (Dough in Bot. Reg. 14, t. 1211), and 
to Pentstemon the remaining species, about 25 in number, in- 
cluding P. centranthifolium {^Chelone centranthifolia, Benth. 
in Hort. Trans. 2d Ser. 1. p. 481.) and P, cordifolium,^ a New 
Californian species collected by Mr. Douglas. 
We now come to the extensive tribe of Gratiolese, embracing a 
great variety of forms, characterised chiefly by the absence of those 
characters which distinguish the several other tribes. The character 
derived from the limb of the corolla is very difficult to observe in 
dried specimens. Indeed, in many cases, the obliquity is so slight 
that the divisions of the corolla are apparently equal to each other, 
in which case the didynamy of the stamina alone distinguishes the 
Gratiolese from the Buddleise and Veronicese. 
Gratiolese are, in the above-mentioned paper, distributed according 
to merely artificial characters. The genera themselves are most of 
them natural, but 1 have not as yet succeeded in collecting them 
* P. cordifolium, caule pruinoso, foliis breviter petiolatis lato-ovatls obtusis 
integerrimis basi cordatis bullatis subtus nervosis, panicula laxa, corollls tubulosis 
profunde bilabiatis, lal)io superiore eniarginato, inferiore trifido, filamento sterili 
barbato, antheris glabris. 
